the developing cognitive substrate of sequential action ...bernhard-hommel.eu/verschoor et al. (in...

RUNNING TITLE: Sequential action control in infancy

The developing cognitive substrate of sequential action control in 9- to 12-month-olds:

Evidence for concurrent activation models

Verschoor1,4, S. A., Paulus4, M., Spapé2, M., Biro1,3, S. & Hommel1, B.

In press

Cognition

1 Leiden University Institute for Psychological Research & Leiden Institute for Brain and Cognition

2 Helsinki Institute for Information Technology

3 Center for Child and Family Studies, Leiden University

4 Developmental psychology unit, LMU München

Keywords: action-effects; eye tracking; pupillometry; goal-directed action; infancy; cognitive

development, sequential action.

Word count: 8211

Development of sequential action control in infancy 2

Abstract

Nine-month-olds start to perform sequential actions. Yet, it remains largely unknown how they

acquire and control such actions. We studied infants’ sequential-action control by employing a

novel gaze-contingent eye tracking paradigm. Infants experienced occulo-motor action

sequences comprising two elementary actions. To contrast chaining, concurrent and integrated

models of sequential-action control, we then selectively activated secondary actions to assess

interactions with the primary actions. Behavioral and pupillometric results suggest 12-month-

olds acquire sequential action without elaborate strategy through exploration. Furthermore, the

inhibitory mechanisms ensuring ordered performance develop between 9 and 12 months of age,

and are best captured by concurrent models.


The developing cognitive substrate of sequential action control in 9- to 12-month-olds:

Evidence for concurrent activation models

Infants are active, goal-directed agents (e.g., McCarty, Clifton, Ashmead, Lee, & Goubet, 2001).

Interestingly, some of the actions they produce can be considered sequential, such as reaching for

a rattle in order to shake it — a rather simple sequence, that comprises two dissociable

components that differ in function and motor demands. Piaget (1936) and others (Claxton, Keen,

& McCarty, 2003; Hauf, 2007; Willatts, 1999; Woodward and Sommerville, 2000; Woodward,

Sommerville, Gerson, Henderson & Buresh, 2009) have stated that true goal-directed action

emerges around 9 months of age when infants begin to be able to organize means-end action

sequences in the service of overarching goals. Yet, the cognitive substrate of early sequential

action control in infants remains completely uncharted territory. The purpose of the current study

is to explore the cognitive mechanism sub-serving sequential action control in infants.

Development of action control in infancy

There are three prerequisites for infants to control sequential action: that they can

represent actions, that they can represent sequential information and that they can combine those

abilities to represent and control sequential action. Let us turn to the first prerequisite. There is

ample evidence that actions are represented in terms of their effects. In his ideomotor theory,

James (1890) states that actions are learned on the fly through sensorimotor exploration; an

automatic mechanism creates bidirectional associations between perceived effects and the

actions producing them (Hommel, 2009; Hommel, Müsseler, Aschersleben, & Prinz, 2001).

These associations bring the actions under voluntary control, enabling the agent to activate the


action by “thinking of” the corresponding effect. The theory can thus account for learning new

actions and new goals.

This idea is typically tested in a two-stage paradigm. Experimenters first let subjects

perform actions that lead to specific effects. After acquisition, they test if exogenously cueing an

effect cues the action that previously caused it (Elsner & Hommel, 2001; Greenwald, 1970). This

approach resulted in demonstrations of bidirectional action-effect acquisition for a wide range of

actions and effects in children (Eenshuistra, Weidema & Hommel, 2004; Kray, Eenshuistra,

Kerstner, Weidema & Hommel, 2006) and adults, suggesting the mechanism responsible to be

fast-acting (Dutzi & Hommel, 2009), automatic (Elsner & Hommel, 2001; Band, Steenbergen,

Ridderinkhof, Falkenstein & Hommel, 2009), implicit (Elsner & Hommel, 2001; Verschoor,

Spapé, Biro & Hommel, 2013), and modulated by the same factors that influence instrumental

learning (Elsner & Hommel, 2004) (for a review on action-effect learning see: Hommel &

Elsner, 2009). Furthermore, action-effects have also been found to be important for action

evaluation (Band, Van Steenbergen, Ridderinkhof, Falkenstein & Hommel, 2009; Verschoor et

al., 2013).

Until recently, research on the importance of action effects for infants mainly focused on

third-person action interpretation (e.g., Biro & Leslie, 2007; Hauf, 2007; Kiraly, Jovanovic,

Prinz, Aschersleben, & Gergely, 2003; Paulus, 2012; Paulus, Hunnius, & Bekkering, 2013;

Woodward, 1998, for a review, see: Hauf, 2007; Kiraly, Jovanovic, Prinz, Aschersleben, &

Gergely, 2003) and imitation (Hauf & Aschersleben, 2008; Klein, Hauf & Aschersleben, 2006;

for a review see: Elsner 2007; Paulus, 2014). Such findings are corroborative in view of the

upsurge of theories stressing similar representations for first- and third-person action (e.g. Baker,

Saxe & Tenenbaum, 2009; Fabbri-Destro & Rizzolatti, 2008; Melzoff, 2006, 2007; Tomasello,


1999). Interestingly, increased model- to self- similarity aids imitation (Shimpi, Akhtar &

Moore, 2013). Yet given their focus on action understanding, such studies tell us little about the

function action effects have for the development of action control in infancy.

Direct evidence regarding action-effect learning was recently obtained from first-person

paradigms similar to that of Elsner and Hommel (2001). Verschoor, Spapé, Biro and Hommel

(2013) showed that 7-month-olds use action effects for first-person action monitoring. By eight

months, infants show motor resonance when listening to previously self-produced action-related

sounds (Paulus, Hinnius, Elk & Beckering, 2011). The youngest infants showing evidence for

reversing bidirectional action effects for action control are 9-month-olds (Verschoor, Weidema,

Biro & Hommel, 2010). Comparable results were found in 12- (Verschoor et al., 2013), and 18-

month-olds (Verschoor et al., 2010). Additionally 6-, 8- (Wang et al., 2012) and 10-month-olds

(Kenward, 2010) anticipate action outcomes. Taken together these studies illustrate that 7-

month-olds represent and monitor first- and third-person action in terms of action effects, while

9-month-olds additionally use action effects for action control.

Representing sequential information in infancy

Another prerequisite for representing sequential action is the ability to encode sequential

information. Infants can register whether items are consistent with familiarized deterministic or

probabilistic sequences (Romberg & Saffran, 2013). For instance, infants are susceptible to

sequential grammar information in speech from birth (Gervain, Berent, & Werker, 2012;

Teinonen, Fellmann, Näätänen, Alku & Huotilainen, 2009), 3-month-olds are susceptible to

spatiotemporal (Wentworth Hait & Hood, 2002) and audio-visual sequences (Lewkowicz, 2008)

and 8-month-olds to analogous information in artificial sound (Marcus, Fernandes & Johnson,

2007). Studies like these suggest an implicit, early-appearing, domain-general statistical


information-acquisition mechanism for sequential information (e.g. Kim, Seitz, Feenstra &

Shams, 2009; Kirkham, Slemmer & Johnson, 2002; Marcovitch & Lewkowicz, 2009) thought to

sub-serve action- and language-segmentation (e.g. Baldwin, Andersson, Saffran & Meyer, 2008;

Saffran, Johnson, Aslin & Newport, 1999). Nonetheless these studies leave open whether infants

encode ordinal information among sequence elements. Indeed, Violation Of Expectation (VOE)

research suggests that while 4-month-old infants encode statistical sequential properties, they

cannot code the invariant order of sequences (Lewkowicz & Berent, 2009). This ability emerges

during the second half of the first year (Brannon, 2002; Picozzi, de Hevia, Girelli & Macchi-

Cassia, 2010; Suanda, Tompson & Brannon, 2008).

Sequential action representation in infancy

The reviewed literature shows that the first two prerequisites for infants’ representation of

sequential action emerge around 9 months. Yet, the question remains whether they can actually

combine these abilities to represent and control action sequences. Indirect evidence comes from

research that suggests infants are able to interpret third-person sequential actions. Evaluating

such actions requires them to be parsed in order to perceive overall syntax and ultimately their

goal (Conway and Christiansen, 2001; Lewkowicz, 2004; Baldwin, Baird, Saylor, & Clark

2001). VOE studies report that around the age of 6 months infants start to evaluate the efficiency

of sequential actions (Biro, Verschoor & Coenen, 2011; Csibra, 2008; Gergely & Csibra, 2003;

Verschoor & Biro, 2012) and causality towards their goals (Baillargeon, Graber, DeVos &

Black, 1990; Woodward & Sommerville, 2000). Olofson and Baldwin (2011) found that 10-

month-olds take into account the kinematics of an observed reaching motion to judge whether it

is part of a familiar action sequence. Yet, Paulus, Hunnius, and Bekkering (2011) showed that

20-, but not 14-month-old infants use such information to predict goals. Additionally,


Gredebäck, Stasiewicz, Falck-Ytter, Rosander and von Hofsten, (2009) showed that 14- but not

10-month-olds’ predictive eye movements are influenced by the models later intention with the

object. Moreover, infants use social context to bind actions of two collaborating actors into

action sequences for goal evaluation (Henderson & Woodward, 2011; Henderson, Wang, Matz

& Woodward, 2013) and goal prediction (Fawcett & Gredebäck, 2013). Although these studies

provide evidence that infants have some understanding of others’ sequential action, they do not

reveal the cognitive mechanisms underlying infants’ control of their own sequential action

Turning to infants’ own action control, studies on (deferred) imitation of enabling action

sequences (sequences in which one action is temporally prior to and necessary for a subsequent

action) report that only a subset of 9-month-olds can (immediately) reproduce such sequences

under ideal circumstances (e.g., Bauer, Wiebe, Waters & Bangston, 2001; Carver & Bauer, 1999,

2001). Adding salient action-effects to separate action steps increases performance (Elsner, Hauf,

& Aschersleben, 2007). However, production of sequential action is in itself not enough to

evince infants’ sequential action control, since subsequent actions may simply be subsequent. In

enabling sequences, stimulus enhancement could externally trigger such sequences. Indeed, an

advantage for imitating enabling- over arbitrarily-ordered actions is reported (e.g. Barr & Hayne,

1996; Bauer, Hertsgaard, & Wewerka , 1995; Mandler & McDonough, 1995, for a review see:

van den Broek, 1997). Earliest evidence for imitation of arbitrarily-ordered action sequences is

reported for 16-month-olds (Bauer, Hertsgaard, Dropik & Daly, 1998).

Advance planning would make a stronger point for sequential action control. Claxton,

Keen and McCarty (2003) reported that 10-month-olds plan the kinematics of reaching

depending on subsequent intentions. Furthermore, McCarty, Clifton and Collard (1999) showed

that 19- but not 14-month-olds inhibit reaching for an object with their dominant hand when this


is inefficient towards an overarching goal (see Cox & Smitsman (2006) for a conceptually

similar result in 3-year-olds). Both McCarty et al.’s (1999) and Cox and Smitsman’s (2006) tasks

depend on inhibition of pre-potent responses and suggest inhibition is important for sequential

action planning (for a review see McCormack & Atance, 2011). Likewise, the disadvantage for

reproducing arbitrarily-ordered action sequences seems to come from an increased need to

temporally organize such sequences (Bauer, Hertsgaard, Dropik & Daly, 1998), which many

theorists hypothesize inhibition to be crucial for (e.g. Constantinidis, Williams & Goldman-

Rakic, 2002; Norman & Shallice, 1986).

To sum up, the studies mentioned above suggest the rudimentary abilities to learn from

and interpret third-person sequential action, as well as the abilities to plan and control first-

person sequential action emerge by the end of the first year. The studies further suggest that

temporal organization, action effects and goals are important sources of information, yet they

leave open how such information is used for integrating action steps into coherent sequences. In

the current study we will attempt to clarify the cognitive mechanism responsible for this feat.

Models of sequential action representation

As there is little specific developmental literature on the subject, we turn to general

psychological theories on sequential action control. Through the years many influential

theoretical incarnations of sequential action representation have been conceived (de Kleijn,

Kachergis, & Hommel, 2014). All of these theories hold that sequential actions consist of

elementary actions that are somehow combined into sequences, as suggested the observation that

the speed of sequence-initiation increases with the number of elements therein (e.g., Henry &

Rogers, 1960; Rosenbaum, 1987). The theories can be distinguished into three ontological types


that differ with respect to the representations action control operates on. We refer to them as

chaining, concurrent, and integrative theories of sequential action control (see Figure 1).

Chaining theories stress that elementary actions are selected and combined through association

processes. Concurrent (Hebbian) theories focus on competitive processes that account for the

orderly production of an action sequence. Integrated approaches highlight crosstalk between

elementary actions resulting in chunked actions.


=== FIGURE 1 ===

Goal

A1 A2

Goal

A1 A2

Goal

A1 A2

A B C

Figure 1. Models of sequential action. Schematic representation of activation in A:

Chaining models of sequential action, activation cascades forward through the different

elementary actions, B: Concurrent models of sequential action, all elementary actions are

activated simultaniously whereafter competion through inhibition ensures the correct order of

execution, C: Integrated models of sequential action, the sequence of actions has been integrated

into a new elementary action.


The prototypical theory of sequential action is James’ (1890) chaining theory. It holds

that elementary actions can be chained by sequentially activating the anticipated effect of each

element. With practice, the sensory effect of each elementary action becomes associated with the

next elementary action through stimulus-response learning, thereby eliminating the need for

sequential activation. The model thus effectively reduces sequential action representation to a

combination of ideomotor and stimulus-response learning. Furthermore, James’ theory can

account for the finding that infants better encode enabling- than arbitrarily ordered action

sequences (Barr & Hayne, 1996; Bauer, Hertsgaard, & Wewerka , 1995; Mandler &

McDonough, 1995), since the proposed feedback dependent effect-response learning he proposes

is aided by stimulus enhancement in such sequences. Although James’ theory is temptingly

simple, it has a number of important drawbacks resulting in additions to the model. Hull (1931)

pointed out that to stay goal-directed and flexible during performance, the representation of the

end state should remain active during execution to compare the actual to the expected outcome.

Hull thus introduces hierarchy into the representation by proposing continuous activation of an

overarching goal. Secondly, in the conception of James (1890) the second action of a sequence is

cued by the sensory effect of the first, suggesting sequential action to rely on sensory feedback.

Yet, empirical evidence suggests feedback mechanisms to be too slow to account for the speed of

practiced sequential action (e.g. Sternberg, Monsell, Knoll & Wright, 1978). Greenwald (1970)

suggested that instead of the sensory effects, the anticipated action effects of the preceding action

are associated to those of the next action. This enables the initiation of the sequence by

anticipating its end effect. However the model does not specify how the end effect activates the

sequence instead of just the final elementary action.


An important criticism on chaining models is that they seem to imply that elementary

actions are equally associated with preceding and subsequent actions, making orderly

performance of sequences impossible. In other words, chaining models of sequential action

assume, but fail to describe how activation moves forward through the sequence. This lack of

temporal dynamics in chaining models resulted in the emergence of a second ontological class of

theories, the concurrent activation theories. Estes (1972) suggested an initial concurrent

activation of all elementary actions by a superseding goal. Thereafter, temporal inhibitory

processes ensure that activation moves forward through the sequence. To guarantee such forward

flow he introduced inhibitory links flowing from each element to the next and secondary self-

inhibition for completed elements (e.g., Henson, 1998, but also James, 1890), equivalent to

inhibition of return in visual attention (Posner & Cohen, 1984; Houghton & Tipper, 1996). The

concurrent model can account for the empirical finding of more prospective than retrospective

intrusion errors (Dell, Burger, and Svec, 1997; Lindenberger, Rünger & Frensch, 2000;

Rumelhart & Norman, 1982, for a review see: Houghton & Tipper, 1996). Concurrent models of

sequential action representation thus utilize inhibition processes which are implicated in studies

on action planning in infants (McCarty et al., 1999; Cox & Smitsman’s, 2006).

The third class of theories, integrative theories of sequential action control, does not

presuppose that action elements remain independent when combined into sequence. Such

theories state that through practice elementary actions can be integrated into one common action

plan or “chunk”, implying considerable crosstalk between the elementary actions (Miller, 1956;

Sakai, Hikosaka, & Nakamura, 2004). Their main support comes from studies that find

reductions in the sequence-length effect by extensive practice (e.g. Klapp 1995). Chunking of the

elementary actions is possible by relating the sequences to internal or external context thus


creating a unique identifying criterion for the associations (Hull, 1931). This Integration process

can account for crosstalk between elementary actions and thus explains end-state comfort effects

(Rosenbaum et al., 1990) as found in infant studies (Claxton, Keen and McCarty, 2003; Cox &

Smitsman, 2006; McCarty; Clifton & Collard, 1999).

Experimental approach

Chaining, concurrent and integrated models generate different predictions with regard to

the spreading of activation from one sequence element to another. Consider a sequence of two

actions, with element R1 preceding element R2. All models imply that priming or otherwise

activating R1 might spread activation to R2, but they differ regarding their predictions when R2

is primed/activated. James’ (1890) chaining model would not predict that priming R2 leads to

activation of R1, since the sequence is assumed to be represented by unidirectional effect-

response bindings (R1àR2). Greenwald’s (1970) version would predict the spreading of

activation from R2 to R1, as sequences are represented by associations formed between the

effects of the elementary actions. Conversely, concurrent activation models would predict that

activating R2 leads to the inhibition of R1, as activation is allowed to spread in forward direction

only and backward connections are inhibitory. Finally, integrated models would predict that

activating one element would activate the representation of the entire sequence, including R1.

The aim of the present study was to pit these different predictions against each other.

In the current study we were not only interested in the cognitive substrate of sequential

action control, but also in the development thereof. Given findings that infants develop the

ability for sequential action control around the end of the first year of life (e.g., Claxton et al.,

2003), we hypothesized to find a developmental change in the cognitive substrate of sequential

action control between 9 and 12 months of age. This line of reasoning is also supported by


findings that the prerequisites for the ability seem to emerge in this interval. In 9-month-olds the

ability to represent sequential information and action is rudimentary at best. Nonetheless, and

crucial to the experimental logic, 9-month-olds represent actions in terms of their effects.

To tackle our questions regarding sequential-action control, we modified a recently

developed gaze-contingent eye-tracking paradigm that assessed action-effect learning in infants

and adults (Verschoor et al., 2013). This paradigm, conceptually identical to that of Elsner and

Hommel (2001), overcomes problems arising due to limited motor control in infants (Verschoor

et al., 2013; Wang et al., 2012). Verschoor and colleagues (2013) first let subjects perform

actions that lead to specific effects. After acquisition, they tested whether exogenously cueing

the effects primes the action that previously caused it. The paradigm uses eye movements which

infants can accurately control from 4 months of age (Scerif et al., 2005) and which can be

considered goal-directed (Gredebäck & Melinder, 2010; Falck-Ytter, Gredebäck, & von Hofsten,

2006; Senju & Csibra, 2008). The paradigm records Reaction Times (RTs) and Response

Frequencies (RFs). The study of Verschoor and collegues (2013) and other recent studies that

demonstrated saccade-effect learning in adults (Huestegge & Kreutzfeldt, 2012; Herwig &

Horstmann, 2011), showed shorter RTs for responses congruent with the previously acquired

action-effect association (Verschoor et al., 2013). There is strong evidence that RT and RF differ

in their sensitivity to congruency effects depending on age. In 9- and 12-month-olds RT is a

sensitive measure (Verschoor et al., 2010; Verschoor et al., 2013), while in 18-month-olds RF

additionally diagnoses congruency effects (Verschoor et al., 2010). The paradigm concurrently

records Task-Evoked Pupillary Responses (TEPRs). The use of TERPs is relatively new in

developmental research (Falck-Ytter, 2008; Jackson & Sirois, 2009; Laeng, Sirois & Gredebäck,

2012; Verschoor et al., 2013). TERPs indicate motivational phenomena such as increased arousal


(Bradley, Miccoli, Escrig & Lang, 2008; Laeng & Falkenberg, 2007), attention allocation (e.g.

Hess & Polt, 1960), cognitive load (Kahneman & Beatty, 1966) and mental effort (Kahneman,

1973; Hess & Polt, 1964). Whatever the exact interpretation of the measure, it enables us to

contrast acquisition contingent vs. non-contingent responses since all interpretations suggested

that dilations should be larger for actions requiring more processing. Furthermore, pupil TERPs

are sensitive to congruency in all ages, showing lesser dilation during congruent action

(Verschoor et al., 2013). Thus, given that we tested 9- and 12-month-olds, we mainly expected

congruency effects on RTs and TERPs.

In previous studies (that all used single-component actions), the definition of congruency

was straightforward: participants were exposed to two action-effect contingencies during

acquisition, in which responses Ra and Rb were followed by action effects Ea and Eb (Ra à Ea; Rb

à Eb). Performing action Ra in response to (or as a result of being primed by) effect Ea in the test

phase (Ea à Ra) would be considered congruent, while performing the same action in response

to Eb (Eb à Ra) would be considered incongruent.

Introducing sequences that consist of two action components (components 1 and 2)

renders the definition somewhat more complicated (see Table 1). Our participants were exposed

to two sequences of actions (A and B) and their effects during acquisition: R1A à E1A à R2A à

E2a and R1B à E1B à R2B à E2B. In the test phase, we presented the action effect of one of the

second action components (E2A or E2B) and we tested whether this would affect processes

related to the first action components (R1A and R1B). If infants represent the experienced action

sequences as a unity, cueing the effect of the second element (E2A or E2B) could affect the

activation of the first action elements (R1A or R1B). The pairings of effect E2A and action R1A


(E2A à R1A) or of effect E2B and action R1B (E2B à R1B) were considered congruent, and the

pairings of E2A and action R1B (E2A à R1B) or of E2b and action R1A (E2B à R1A) incongruent.


=== TABLE 1 ===

Table 1. The action sequences learned during acquisition and the congruent and

incongruent responses during test.

ACQUISITION TEST

SEQUENTIAL ACTION CONGRUENT INCONGRUENT

R1A à E1A à R2A à E2a

And

R1B à E1B à R2B à E2B

E2A à R1A

And

E2B à R1B

E2A à R1B

And

E2B à R1A


Finding any difference depending on congruency would provide evidence for a coherent

cognitive representation of sequential action in infants. Crucially, the direction of the effect

would speak to the internal structure of that representation: While chaining and integrative

models would lead one to expect facilitation (shorter latencies and smaller pupil dilations) in

congruent responses, concurrent activation models would predict the opposite.


METHODS

Subjects

Two age groups were tested: 14 9-month-olds (mean: 8.94 months, SD= .37, SE=.9, 5

female) and 16 12-month-olds (mean: 11.99 months, SD= .42, SE=.10, 9 female), another 4 9-

and 7 12-month-olds were excluded for not meeting the criterion for the minimal amount of test

trials. They were recruited through the municipality and received small gifts as compensation.

An informed consent and a questionnaire regarding general health and development were

obtained. The infants were all healthy full-term and without pre- or perinatal complications.

Test environment and apparatus

During the experiment the infants sat in a specially designed, stimulus-poor booth on the

lap of their caretaker, who was seated in front of the eye-tracker apparatus. The distance between

eyes and apparatus was approximately 70 centimeters (the screen’s viewing angle was 34.1° by

21.8°). The behavior of the infants was monitored online by the experimenter from a separate

control room by means of a camera located above the apparatus. A 17 inch TFT-screen (1280 x

1024 pixels), equipped with an integrated Tobii T120 eye-tracker operating at 60 Hz, was used

for visual and auditory data presentation, and data collection. The Tobii T120 has an optimal

accuracy of .5° and allows for a certain amount of head movement by the subjects

(30x22x30cm). It recorded gaze direction and pupil-size. Stimulus presentation was controlled

by a PC running E-prime® software (Schneider, Eschman & Zuccolotto, 2002).

Stimuli

The visual stimuli used were as follows (see Figure 2). The background color of the

screen was grey. The fixation point was a brightly colored dot with a superimposed line drawing


(4.3° by 4.3°). To keep infants interested, the color of the dot changed randomly from trial to

trial (selected from eight colors) and the line drawing was randomly selected (without

replacement) from a selection of 50 drawings (Snodgrass & Vanderwart, 1980). As Response

Areas (RA’s), we used 100 grayscale pictures from the “Nottingham scans” faces database,

(http://pics.psych.stir.ac.uk), displaying emotionally neutral frontal faces of 50 men and 50

women. Faces were chosen to elicit spontaneous saccades as they are known to attract infants’

attention (Goren, Sarty & Wu, 1975; Johnson, Dziurawiec, Ellis & Morton, 1991). To maximize

the chance of finding an effect, the faces looked at the participant, since Sato and Itakura (2013)

showed that eye contact enhances action-effect binding. We used two pairs of 200ms effect

sounds which were equalized on loudness, “tring” and “piew” (Verschoor et al.,2013, 2012) and

complex high- and low- note sound waves of 1574- and 776-Hz.

Procedure

Infants were tested at a time when they were likely to be alert. Prior to the experiment the

caretakers were instructed not to move after calibration and gently hold the infant in order to

maintain eye-tracker alignment, and to entertain the infant during the 1-min interruption between

calibration and the experiment. The eye-tracker was calibrated using a 9-point calibration

consisting of a small animation. The calibration was accepted with a minimum of eight points

acquired. The experimenter could play an attention-grabbing sound during the experiment. If this

no longer worked caretakers were encouraged to direct the infant’s attention to the middle of the

screen by pointing. Lighting conditions were kept constant. Furthermore, luminance levels were

controlled for by presenting the stimuli in a random fashion. After completion an explanation of

the experiment was provided.


Acquisition phase

The experiment began with an acquisition-phase of 36 trials (see Figure 2). If during the

acquisition phase the subject showed declining attention, the acquisition phase could be

shortened (minimum number of acquisition trials was set at 24). In each trial participants could

freely choose to perform one of two saccade sequences (R1A à E1A à R2A à E2a or R1B à E1B à

R2B à E2B). Each saccade sequence consisted of two distinct actions, first one to the left or right

(R1A or R1B) whereafter an up- or downward action followed (R2A or R2B, depending on the

mapping assigned). Each saccade was followed by an effect-sound which was consistently

designated to left-, right-, up- and downward Response Areas (RA’s).

A trial started with the fixation dot. The dot disappeared after fixation on it for an interval

that varied (to remove any bias or habituation caused by fixed intervals), between 150- and 350-

ms. After disappearance, photographs of two different faces (randomly selected without

replacement from 100 pictures) appeared to the left and right. The faces served as Response

Area’s (RA’s). The 5.3° by 5.3° pictures appeared at 9.7° to center. To avoid perseverance to

either side across trials the images pulsated. One of them started shrinking to 4.1° while the other

started growing to 6.5° (side shrinking was randomized); one cycle from intermediate size to

small, to intermediate, to large and back to intermediate, took 2 s.

When a saccade towards one of the faces was detected it stopped pulsating and the other

face disappeared. Depending on the targeted side, one of two distinct 200ms effect-sounds

(“tring” or “piew”) was presented (E1A or E1B, the mapping was balanced across participants).

RA’s were defined as the maximum size of the pulsating images: 6.5° by 6.5°. A saccadic

response was defined as eye movement (minimally 4.3°) into the left or right response area.


Immediately after the effect the current face disappeared and reappeared 7.8° above or below

that location (depending on the mapping) in the same dimension and continued to pulsate serving

as RA again (again defined as the maximum size of the image). Upon detection of a saccade to

that location (minimal 1.3°), one of two distinct 200ms effect-sounds (E2A or E2B, “high note” or

“low note”) was presented (the mapping was balanced across participants). RTs were defined as

the time interval between disappearance of the fixation dot and detection of a saccade in the

secondary RA. The maximum allowable RT was 2000ms; if by then no response was detected,

the trial was repeated. After each trial, an inter-trial-interval of 500ms was used.

Test phase

The test phase of 32 trials followed directly afterwards (see Figure 2). The minimum

number of test trials to enter analysis was 22. A trial started with the fixation dot as during

acquisition. However, after fixation (fixation time identical to acquisition), the dot remained on

display for 200ms during which an effect-sound was presented that was previously triggered by

one of the two secondary eye-movements (E2A or E2B). Thereafter the dot disappeared. Then, two

identical 5.3° by 5.3° images of the same face (randomly selected without replacement) appeared

9.7° to the left and right of the screen center serving as RAs. The two images were identical to

minimize gaze preference. To further reduce bias the faces pulsated in synchrony, meaning that

they either both grew or shrank (randomized and with the same motion parameters as during

acquisition). Again, the images were expected to evoke saccades (R1A or R2B). The question of

interest was whether the direction of these saccades (R1A or R2B) would be biased by the tones

(E2A or E2B). Except for absence of auditory effects after the saccades, the remaining procedure

was as during acquisition.


=== FIGURE 2 ===

Test phaseAcquisition phaseT1

T2

T3

T4

TIME

1 or 2

3 or 4

TIMET1

T2

T3

3 or 4


T2

T3

T4

TIME

1 or 2

3 or 4

TIMET1

T2

T3

3 or 4

Figure 2.

Acquisition trial.: Each trial starts with an intertrial interval of 500 ms. T1: A fixation

dot is displayed at screen center. T2: After successful fixation, faces appear at either side of the

screen where they started to pulsate. T3: Depending on the saccade target, the face at the other

side disappears and an effect sound is played for 200 ms. T4: Depending on which side was

chosen the face moves up or down whereafter a second saccade is made and a second sound

effect is played for 200 ms.

Test trial. Each trial starts with an intertrial interval of 500 ms. T1: A fixation dot is

displayed at screen center. After succesful fixation one of the secondary effects from the

acquisition is played. T2: The dot diasppears whereafter the same face appears on both sides. T3:

The participant freely chooses where to saccade.


Data acquisition

E-prime® 1.2 (Psychology Software Tools, Sharpsburg, PA) was used to collect RTs, the

number of left and right responses and congruent and incongruent responses during test. The

gaze- data files Tobii produced were imported into BrainVision Analyzer 1.05 (BrainProducts

GmbH, Gilching, Germany) to analyse gaze position and pupillary data. Depending on analysis,

segments were created from 2000ms before the presentation of the sound onset or RT, to 8000ms

after, while allowing overlapping segments. Responses were sorted on congruency of the

response and stimulus- and response-locked functions were averaged (Verschoor et al., 2013).

Following Bradley et al. (2008), pupil-diameter measurement began after the initial pupil reflex

caused by the fixation stimulus. Visual inspection showed it to end around 500ms after effect

presentation (see Figure 4) (see also Verschoor et al., 2013). Dilations were calculated as the

percentage of dilation relative to the baseline to make the data comparable across age groups.

The percentage of trials rejected due to erroneous data points (leaving 29 valid trials on average)

did not differ across age groups, p > .8. Dilations of both eyes were averaged to reduce noise.

Artifacts and blinks detected by the eye-tracker were corrected using a linear interpolation

algorithm, after which a 10 Hz low-pass filter was applied (c.f., Hupe, Lamirel & Lorenceau,

2009). Further artifact rejection was done using a threshold based approach, including those

segments with pupil sizes between 1 and 5 mm, and a maximum change in pupil size of .03mm

in 17ms. Gaze data were recorded in pixel coordinates, averaged between eyes and filtered using

a 10Hz low-pass filter.

Given that the acquisition of action-effect associations is sensitive to the same factors as

stimulus-response learning (Elsner & Hommel, 2004), the Number Of Completed Acquisition


Trials (NOCAT) was taken as an individual measure of action-effect learning. The Mean

Acquisition Reaction Time (MART) was taken as an individual measure for general speed and

activity. Both NOCAT and MART variables were used as covariates in the analyses when

appropriate.

RESULTS AND DISCUSSION

Acquisition phase

First we tested for age group differences in dependent variables collected during

acquisition to ensure that the learning experiences of the age groups were comparable (see Table

2). All ANOVA’s were performed with age group as a between-subjects factor. There were no

effects for the percentage of completed acquisition trials (p >.5), mean RT (p >.5), or the

percentage of right vs. left responses (p >.2) or upward vs. downward responses (p >.2). Two

reliable effects were obtained for RTs. Firstly, horizontal response location interacted with age

group, F(1,25)=6.25, p= .019, η2p=.20. Separate analyses showed no main effect in 9-month-

olds (RT-left=999ms, RT-right=104 6ms) and a tendency toward faster right-ward responses in

12-month-olds, F(1,12)=3.84, p=.074, η2p=.24 (RT-right=982ms, RT-left=1089ms). Secondly,

vertical response location interacted with age group, F(1,25)=4.63, p=.04, η2p=.16. Separate

analyses showed no main effect in 9-month-olds (RT-up=1008ms, RT-down=1037ms) and a

tendency toward faster downward responses in 12-month-olds, F(1,12)=3.82, p=.07, η2p=.24

(RT-up=1089, RT-left=982ms). We also performed a repeated measures ANOVA on RT’s with

Time (dividing the responses in three equal bins) and found no effect (p >.13). Lastly, we

performed a repeated measures ANOVA on the partial RTs of the primary action to test if

contraction vs. expansion had an effect on these partial RTs. We found a significant effect,


F(1,28)=175, p <.001, η2p=.86, indicating responses toward contracting pictures were slower

(partial RT-contracting=603ms, partial RT-expanding=428ms).


=== TABLE 2 ===

Table 2. Mean scores of acquisition phase (standard deviation in brackets).

AGE GROUP SCORES

Percentage of completed acquisition

trials

Percentage of left responses RT in ms RT

left RT

right RT up RT down

9-month-olds 92.9 (11)

42.8 (38)

1007 (81)

999 (107)

1046 (90)

1008 (116)

1037 (83)

12-month-olds 92.0 (14)

38.5 (40)

1023 (83)

1089 (167)

982 (73)

982 (78)

1089 (164)


We concluded that the learning experiences were comparable across age groups. The

interaction of horizontal response location and age on RTs might reflect the fluctuating

emergence of general right-side preference during the first year (Corbetta & Thelen 1999;

Michel, 1998), which also affects infants’ eye movements (Cohen, 1972). An orthogonal effect

may be reflected in our analysis of upward vs. downward RTs. However, little is known about

such preferences. Additionally, we found that infants responded faster toward expanding

pictures. This effect probably reflects automatic attentional processes to avoid collisions (e.g.,

Kayed & Van der Meer, 2000; Van Hof, Van der Kamp & Savelsbergh, 2006). Importantly,

these observations are not detrimental to our research question since both age groups received

approximately the same amount of training for all combinations of response locations.

Test phase

All ANOVA’s were performed with age group as a between-subjects factor. There was

no effect on the percentage of completed test trials (p >.4).

Response frequency

Overall, participants looked more often (64%) to the right than left side, F(1,28)=9.00,

p=.02, η2p=.19, but the effect did not interact with age. More important for our purposes,

ANOVA’s with congruency as within-subjects factor were not significant, adding MART,

NOCAT or both as covariates didn’t change this ( p’s >.2). We concluded that, if infants control

sequential actions, this does not seem to affect the probability to choose a particular sequence.


Reaction times

There were no reliable effects with regard to overall RT (p > .6), left vs. right response

location (p >.3) (see Table 3) or inter-trial interval, (p > .5) (which we analyzed because the test-

phase was self-paced). More important for our purpose, an ANOVA with congruency (see Table

1 for mapping details) as within-subjects factor, revealed a significant effect indicating 29ms-

slower responses for congruent trials, F(1,28)=4.15, p=.05, η2p=.13; the interaction with age

was not significant (p >.3). Although the statistics did not necessitate further exploration, given

our directed hypothesis about age effects, we looked at both age groups separately. In the 9-

month-olds the effect was not significant (p >.4) while in the 12-month-olds it was F(1,15)=5.47,

p =.03, η2p=.27. A non-parametric Wilcoxon signed rank test confirmed these results (9-month-

olds: Z=-1.57, p=0.88, 6 of 14 infants showed the pattern, 12-month-olds: Z=-2.02, p=0.04, 12 of

16 infants showed the pattern). However, adding NOCAT as a covariate into the separate

ANOVA for the 9-month-olds resulted in a significant effect (F(1,12)=4.96, p=.05, η2p=.29).


=== TABLE 3 ===

Table 3. Mean frequency and RT scores of test phase (standard deviation in brackets).

AGEGROUP SCORES

Percentage completed test trials

Percentage left

responses

Percentage congruent responses

ITI ms

RT ms

RT Congruent

ms

RT Incongruent

ms


43.2 (38)

47.5 (8)

1637 (323)

431 (90)

440 (104)

424 (91)

12-month- olds

96.5 (10)

29.0 (21)

49.3 (7)

1563 (393)

447 (74)

468 (83)

425 (83)


Our main finding is: cueing of the secondary action of the action sequence interfered with

executing its first, as evidenced by the longer RT’s for congruent responses in the 12-month-olds

(see Table 1 for mapping details). Results were less clear in the 9-month-olds. The signed rank

test did not show significant results while using Number of Completed Acquisition Trials

(NOCAT) as a covariate in the RT analysis resulted in a significant effect in 9-month-olds. This

suggests that the NOCAT was an important factor for the strength of the effect in this age group

whereas the 12-month-olds showed a ceiling effect for the NOCAT necessary for the uptake of

the sequential action. The fact that we found an effect can be considered as evidence that 12-

month-olds control sequential action. Performing two consecutive actions is sufficient to

integrate them into a coherent representation. Twelve-month-olds apparently represent action

sequences in a format that allows for interactions between the codes of their individual elements

(which excludes fully symbolic formats). Moreover, our findings provide specific support for

concurrent activation theories, as only these would predict interference. Furthermore our findings

suggest sequential action control is developing in 9-month-olds.


=== FIGURE 3 ===

Figure 3. Mean reaction times (+SE) for 9-month-olds (N=14) and 12-month-olds (N =

16) in congruent and incongruent test trials.


Pupil dilation

To accommodate for the variable RTs across age groups and conditions, we considered

both stimulus-locked and response-locked Task-Evoked Pupillary Responses (TEPR’s). The

stimulus-locked analysis for congruent and incongruent (see Table 1 for mapping details)

responses TERPs used a 500ms pre-effect baseline (Beatty & Lucero-Wagoner, 2000). A

repeated measures ANOVA on TERPs with congruency as within subjects factor revealed no a

priori effects of congruency on baselines (-500 to 0ms), p’s > .7. Adults’ TEPRs start from 200

to 300ms after stimulus onset and peak in the range of 500ms to 2000ms (Beatty, 1982; Beatty &

Lucero-Wagoner, 2000). We therefore calculated the mean TERPs for congruent and

incongruent responses as the mean percentage of change from baseline to 500-2000ms post

effect onset. An ANOVA with MART as covariate revealed that, overall, participants exhibited

larger relative pupil dilations during congruent responses, F(1,27)=4.12, p=.05, η2p=.13,

independent of age group (p >.7). Since the time window was based on adult findings, which

likely underestimate the pupillary reactions of the slower infants (Verschoor et al., 2013), we

reran the analysis with a 1000-2500ms post effect onset time window. Again, pupil dilations

were significantly larger in congruent trials, F(1,25)=5.03, p=.03, η2p=.16, independently of

age, p >.09 (see Figure 4).


=== FIGURE 4 ===

Figure 4. Relative pupil sizes for congruent and incongruent responses to baseline,

stimulus-locked.


For the response-locked analysis, we calculated the percentage of dilation from a 700-ms

time window starting at saccade onset, to the same 500ms pre-stimulus baseline. An ANOVA

with MART as covariate yielded a tendency for larger relative dilation in congruent trials,

F(1,27)=3.51, p=.07, η2p=.12, while the interaction with age group was not significant, p >.8

(see Figure 5). Adding NOCAT as additional covariate resulted in a significant effect

(F(1,26)=5.48, p=.03, η2p=.17), again without an interaction with age group (p >.9).


=== FIGURE 5 ===


response locked.


Finding larger relative pupil dilations for congruent responses in both stimulus-locked

and response-locked analyses corresponds nicely to the outcome of the RT analysis. Cueing the

second component of an action sequence makes the execution of the first slower and more

effortful.

Gaze position

In the test phase we primed the second action of the two-element sequences carried out in

the acquisition phase by presenting the corresponding action effect (see Table 1 for mapping

details). Activating the second element of the sequence might affect action planning directly.

One of the second elements was an upward movement while the other was a downward

movement. Priming these elements by their effects might induce a vertical bias in the direction of

the cued element. Alternatively: the selection of the primary action results in forward inhibition

of the second. This might induce the opposite bias. To investigate these effects, we analyzed the

mean vertical deviation from the horizontal midline toward the primed action element as a

function of congruency. To do this we collapsed all vertical deviations from horizontal midline

toward the direction cued to one side and divided the data segments according to congruency

from stimulus onset to 650ms thereafter (corresponding to the mean RT plus mean random ITI)

and compared these segments to a 150ms pre-effect baseline (the minimum fixation time before

effect onset).

There were no a priori effects of congruency on baselines, p> .5. An ANOVA with

MART as covariate showed that during congruent responses gaze position deviated vertically

significantly less toward the direction cued by the effect sound, F(1,27)=4.83, p=.04, η2p=.15


(effect size = 22 pixels; see Figure 6) than in incongruent responses, and this effect did not vary

with age, p >.5. We additionally performed separate ANOVAs with MART as covariate testing

congruent- and incongruent- responses against no deviation. The effect was significant for

incongruent responses (F(1,27)=4.70, p=.04, η2p=.15) and did not vary with age ( p=.2), but not

significant for congruent responses (p >.26).


=== FIGURE 6 ===

Figure 6. Vertical distance from midline toward cued direction of gaze position in pixels

for congruent and incongruent responses. Time=0, is the moment the effect starts.


Our gaze position findings show that priming the second action component results in

activation of the vertical component only in incongruent trials. This might be due to competition

between activated components in congruent trials, as concurrent models would hold: the

selection of the primary action results in forward inhibition of the second. This finding provides

further evidence for the concurrent model of sequential action control in infants and highlights its

temporal dynamics. Moreover, the finding is not in accordance with integrative models since no

vertical bias was found for the primary actions in congruent trials.

GENERAL DISCUSSION

The aim of the current study was to examine (the development of) the cognitive substrate

for sequential action control in 9- to 12-month-olds using a novel gaze-contingent paradigm.

Relying on the idea that if two elementary actions are bound together in a sequence, priming the

secondary action component would influence the availability of the primary component, we

presented the infant participants with a two-step action sequence. While chaining and integrative

models would lead one to expect facilitation in congruent responses, concurrent activation

models would predict the opposite. Our major finding is that priming the second action inhibits

the primary action, as indicated by latencies and pupil dilation. Secondly, we found an effect on

gaze position indicating that action control inhibits the second component of an action sequence

whilst preparing the first part of the sequence. Our findings on three different measures suggest

an emerging ability for sequential action control in 9-month-olds that fully develops by the first

birthday, and is best captured by concurrent activation models (Estes, 1972).

From a developmental perspective our findings extend behavioral studies suggesting

infants can control sequential action (e.g., Claxton, Keen & McCarty, 2003; McCarty, Clifton


&Collard, 1999), and studies showing that this ability to be present only under ideal

circumstances in 9-month-olds, or only a subset of subjects of this age group (Bauer, Wiebe,

Waters & Bangston, 2001; Carver & Bauer, 1999, 2001; Elsner, Hauf, & Aschersleben, 2007;

Lukowski et al., 2005 Waters & Bangston, 2001). In addition, they are in accordance to studies

suggesting that during the second half of the first year the ability to encode ordinal information

comes online (Brannon, 2002; Picozzi, de Hevia, Girelli & Macchi-Cassia, 2010; Suanda,

Tompson & Brannon, 2008). The fact that evaluation of third-person sequential action is

apparent significantly earlier in development, in 6- to 7-month-olds (Baillargeon, Graber, DeVos

& Black, 1990; Biro, Verschoor & Coenen, 2011; Csibra, 2008; Gergely & Csibra, 2003;

Verschoor & Biro, 2012), tentatively suggests either a different cognitive substrate or a

dissociation between evaluation and production (e.g., Verschoor et al., 2013).

Furthermore, our findings relate to infant studies (Claxton, Keen & McCarty, 2003;

McCarty, Clifton & Collard, 1999; Cox & Smitsman, 2006) and cognitive theories (e.g.

Botvinick & Plaut, 2004; Constantinidis, Williams & Goldman-Rakic, 2002; Cooper & Shallice,

2006; Estes, 1972; Norman & Shallice, 1986 Rumelhart & Norman, 1982) that implicate

inhibition as an imperative faculty for controlling sequential action. Interestingly, inhibitory

control begins to emerge toward the end of the first year and undergoes rapid development

across the toddler period and into the preschool years, a pattern coinciding with age-related

changes in frontal lobe maturation and connectivity (Diamond, 2002, Diamond et al., 2007,

Luria, 1973; Wolfe & Bell, 2007). This onset around 1 year of age relates to the developmental

timeline revealed by our results and supports our interpretation that inhibitory processes play an

important role in the ontogenesis of sequential action control.


Note that the development of inhibitory capacities has been linked to the development of

time perception itself (Zélanti & Droit-Volet, 2011; Mäntylä, Carelli & Forman, 2007).

Furthermore, in clinical (Barkley, 1997; Gerbing, Ahadi & Patton, 1987; Montare, 1977) and

healthy populations (Foster et al., 2013) tests of inhibition show robust relationships to indices of

timing (-deficiency). Thus the question arises whether the inhibitory mechanisms found are

specific for action control or are general for representing temporal events (Fuster, 1993, 2002;

Norman & Shallice, 1986). The literature reviewed here seems to point to the latter, suggesting

temporal representations in the form of concurrent activation may be a precondition for

sequential-action control. One might thus speculate that very early sequential-action evaluation

(e.g., Verschoor & Biro, 2012) depends on non-ordinal, or non-temporal representations.

Interestingly, our paradigm offers a possibility to address these and related questions in future

research.

Concerning current theories on action control, our results seem to point to limitations in

explanatory power of the current ideomotor theory (Hommel et al., 2001; Shin, Proctor &

Capaldi, 2010) with regard to sequential action control, as this theory would predict activation of

actions by their effects whereas we find inhibition of the primary action. Sidestepping the idea

that inhibition of the primary action is not the same as inhibition of the sequence as a whole, Hull

(1931) pointed out that binding of sequential action is possible by relating the sequences to

internal or external context such as an overarching goal. An interesting question that such

reasoning poses, is what kind of context and how such context may be incorporated in an

overarching goal representation. Indeed action-effect learning can be context-specific (Kiesel &

Hoffmann, 2004) which could accommodate such overarching ideomotor representations of

action sequences. Thus we may not have succeeded in cueing the overarching goal because of


insufficient context in the cue, and might have gotten stuck in the underlying concurrent level of

representation. Indeed, Kiesel and Hoffmann (2004) have shown that the same actions can be

accessed by different effect anticipations. They also claim that that response initiation has to wait

for the anticipation of the effects that trigger the response (see also Kunde, 2003), suggesting it

takes longer to initiate a response if it produces a long effect. Although this theory would also

predict slower initiation for sequential actions, if one thinks of a sequence of (actions and) effects

as a long effect, the theory cannot account for competitive processes our findings suggest. Thus

we suggest ideomotor theory should be enhanced by incorporating overarching- and sequential-

levels of goal representation. Such hierarchical structure might be conceived as either structural

(Cooper & Shallice, 2006) or epiphenomenal (Botvinick & Plaut, 2004) to concurrent activation

models.

Nonetheless, our findings do suggest ideomotor processes play a role in sequential action

since we inhibited the primary and secondary action components by cueing the secondary action

via its effect. Ideomotor processes have indeed been implicated in sequential action (Koch,

Keller & Prinz, 2004). Ziessler (1994, 1998) and Elsner et al. (2007) found that action-effects

play an important role in sequence learning and Stöcker & Hoffmann (2004) found that action

effects facilitate chunking. Furthermore, the activation of the secondary vertical component in

the incongruent trails is direct evidence for ideomotor theory. Thus although the current findings

extend cognitive theories of action control by suggesting that ideomotor theory needs

elaborations to incorporate sequential action (see for a similar point: Herbort & Butz, 2012;

Kachergis, Wyatte, O'Reilly R, De Kleijn, & Hommel 2014), they do not counter the ideomotor

principle itself.


Another theoretical implication of our results is that the repeated successiveness of

actions in the acquisition phase sufficed to bind the actions into a sequence. This raises the

question of what the exact criteria might be for such binding to occur. One could think of several

dimensions for such criteria; our study suggests repetition, temporal closeness and spatial

closeness might play a role. This is a particularly interesting question since its answer might

provide clues as to how the cognitive system generates new action sequences, never performed

before. However, more research is needed to answer such questions in more detail. For now, our

results suggest that concerning infants’ own action control, sequential action can be picked up by

exploration and does not necessarily depend on elaborate abstract or explicit strategies

(Cleeremans & McClellend, 1991) that operate in terms of efficiency (e.g. Gergely & Csibra,

2003) or causality (e.g., Woodward & Sommerville, 2000).

Even though we consider the present findings as a first step towards the understanding of

sequential-action control in infants, further research is needed to explore this model in greater

detail. Although our paradigm produced continuous data which are temporally rich, they

nevertheless should be considered as a snapshot of processes at work in sequential-action

control. Earlier we hypothesized that cueing the secondary effect of the two-step sequence might

have been too poor in contextual information to cue the overall sequence, thus resulting in local

competition effects in an underlying concurrent level of representation. Alternatively, more

dynamic explanations could be considered. For instance, it could be that the inhibition of the first

sequence components was due to temporal differences in the process of activating the individual

action components on the one hand and of the overarching goal representation on the other. It is

well documented that initiating more complex sequential actions takes longer than initiating

simpler actions (Henry & Rogers, 1960; Rosenbaum, 1987). One could thus speculate that


cueing the second action component activated the underlying representation quickly, but it took

more time to activate the overarching goal representation. The eventual activation of this goal

representation could have facilitated both components of the sequence (as proponents of

integration theories might suggest), but that may have taken too long to be picked up by our

measures. As a consequence, the inhibition that our findings point to may reflect an initial state

of a dynamic action-planning process. Another possibility would be that cueing an action

component that is not yet appropriate (as none of the secondary components was a valid action in

the test phase) resulted in the inhibition of not only the first component but of the entire

sequence, perhaps including the goal representation. We cannot exclude that the second

component of each sequence was also inhibited—although the lack of gazing “away” from the

direction cued by the second component suggests that it was not. The current experiment was not

set up to distinguish between these more detailed scenarios.

Other studies have shown end state comfort effects in infants (Claxton, Keen & McCarty,

2003; McCarty, Clifton & Collard, 1999; Cox & Smitsman, 2006) indicative of integrated

representations of sequential action. In the current study we did not find evidence for this model.

Nonetheless, we do not wish to claim that integrated sequential action control cannot occur in

infancy. We would like to stress that the chaining, concurrent and integrated theories of

sequential-action control are by no means mutually exclusive or complete. They posit useful

approximations for understanding sequential actions, yet depending on exact circumstances

relating to practice, content, time pressure and strategy, some models may be more adept than

others at explaining specific empirical phenomenon. In our opinion a future all-encompassing

theory of sequential action control will probably encompass elements of all three classes of

theories. Indeed our results on gaze directions indirectly suggest an influence of the secondary


action on the primary action that was cancelled out by counteracting inhibitory and excitatory

processes. However, this does not diminish the importance of showing that concurrent processes

are at work in infant sequential-action control.

One could question whether our findings are generalizable to other action systems

(manual, postural etc.). Ideomotor theory makes no distinction between effectors and effect

modalities, and ideomotor motivated research does not suggest such a distinction. Furthermore,

saccade-effect learning is now established in adults (Huestegge & Kreutzfeldt, 2012; Herwig &

Horstmann, 2011) and infants (Verschoor et al., 2013) suggesting that the oculomotor system is

controlled in the same way as for instance the system for manual action control. We

acknowledge that saccadic eye movements operate on very short timescales where priming and

inhibition may play a larger role than in other types of actions that involve gross movements of

the body (e.g. reaching and locomotion) that operate on relatively slower timescales. Future

research will have to clarify this issue.

Altogether it remains essential to develop a comprehensive theory of (the development

of) sequential action representation, which specifically addresses the question of how novel

components are integrated into a sequential plan, how the sequence is generated, whether this

requires hierarchical representations and what types of information are incorporated in

overarching goals. We are confident that further modifications of our paradigm will help to

increase insight into the general cognitive mechanisms underlying action planning (e.g., by

cueing the first action component and examining how this affects the availability of the second)

since our synergy of methodology provides various measures (frequency-, RT-, pupillary- and

gaze position measures) that can pick up different dynamic aspects of the planning process.


In conclusion, the current study shows that sequential action can be picked up by

exploration and does not depend on elaborate abstract strategies. Furthermore, the present

findings demonstrate that 12-month-olds are able to construct action plans comprising more than

one element, and use inhibition mechanisms as suggested by concurrent activation models to put

elements into the right temporal order. And lastly, we provide further evidence for the claim that

the ability for sequential-action control develops between 9 and 12 months of age.

DISCLOSURE

The authors declare no competing interests.

ACKNOWLEDGEMENT

This research was supported by the Netherlands Organization for Scientific Research. We

thank the reviewers for their useful suggestions and especially thank Thijs Schrama for technical

support and Henk van Steenbergen for analytical support.

REFERENCES

Baker, C. L., Saxe, R., & Tenenbaum, J. B. (2009). Action understanding as inverse

planning. Cognition, 113, 3, 329–349.

Baldwin, D .A., Baird, J. A., Saylor, M. M., & Clark, M. A. (2001). Infants parse

dynamic action. Child Development, 72, 708-717.

Baldwin, D., Andersson, A., Saffran, J., & Meyer, M. (2008). Segmenting dynamic

human action via statistical structure. Cognition, 106, 1382-1407.

Band, G. P. H., van Steenbergen, H., Ridderinkhof, K. R., Falkenstein, M., & Hommel,

B. (2009). Action-effect negativity: Irrelevant action effects are monitored like relevant

feedback. Biological Psychology, 82, 211-218.


Barkley, R. A. (1997). Behavioral inhibition, sustained attention, and executive functions:

Constructing a unified theory of ADHD. Psychological Bulletin, 121, 65–94.

Barr, R., & Hayne, H. (1996). The effect of event structure on imitation in infancy:

Practice makes perfect? Infant Behavior and Development, 19, 255-259.

Barr, R., Dowden, A., & Hayne, H. (1996). Developmental changes in deferred imitation

by 6- to 24-month-old infants. Infant Behavior & Development, 19, 159–170.

Bauer, P. J., Hertsgaard, L. A., & Wewerka, S. S. (1995). Effects of experience and

reminding on long-term recall in infancy: Remembering not to forget. Journal of Experimental

Child Psychology, 59, 260–298.

Bauer, P. J., Hertsgaard, L. A., Dropik, P., & Daly, B. (1998). When even arbitrary order

becomes important: developments in reliable temporal sequencing of arbitrarily ordered events.

Memory, 6, 165–98.

Bauer, P. J., Wiebe, S. A., Waters, J. M., & Bangston, S. K. (2001). Reexposure breeds

recall: effects of experience on 9-month-olds’ ordered recall. Journal of Experimental Child

Psychology, 80, 174–200.

Beatty, J. (1982). Task-evoked pupillary responses, processing load, and the structure of

processing resources. Psychological Bulletin, 91, 276–292.

Beatty, J., & Lucero-Wagoner, B. (2000). The pupillary system. In J. Caccioppo, L.G.

Tassinary, & G. Berntson (Eds.), The handbook of psychophysiology (2nd ed.) (pp. 142-162).

Cambridge University Press.

Botvinick, M., & Plaut, D. C. (2004). Doing without schema hierarchies: A recurrent

connectionist approach to normal and impaired routine sequential action. Psychological Review,

111, 395–429.


Bradley, M. M., Miccoli, L., Escrig, M. A., & Lang, P. J. (2008). The pupil as a measure

of emotional arousal and autonomic activation. Psychophysiology, 45, 602-607.

Brannon, E. M. (2002). The development of ordinal numerical knowledge in infancy.

Cognition, 83, 223–240.

Biro, S., & Leslie, A. (2007). Infants' perception of goal-directed actions. Development

through cues-based bootstrapping. Developmental Science, 10, 379-398.

Biro, S., Verschoor, S. & Coenen, L. (2011). Evidence for a unitary goal-concept in 12

months old infants. Developmental Science, 14(6), 1255–1260.

Claxton, L.J., Keen, R., & McCarty, M. E. (2003). Evidence of motor planning in infant

reaching behavior. Psychological Science, 14, 354-356.

Botvinick, M., & Plaut, D. C. (2004). Doing without schema hierarchies: A recurrent

connectionist approach to routine sequential action and its pathologies. Psychological Review,

111, 395–429.

Carver, L. J. & Bauer P. J. (1999). When the event is more than the sum of its parts: 9-

month-olds' long-term ordered recall. Memory, 7, 147-174.

Carver, L. J., & Bauer, P. J. (2001). The dawning of a past: the emergence of long-term

explicit memory in infancy. Journal of Experimental Psychology, 130, 726–745.

Cohen, L. B. (1972). Attention-getting and attention-holding processes of infant visual

preferences. Child Development, 43, 869-879.

Constantinidis, C., Williams, G. V. & Goldman-Rakic, P. S. (2002). A role for inhibition

in shaping the temporal flow of information in prefrontal cortex. Nature Neuroscience, 5, 175–

180.


Conway C. M., & Christiansen M. H. (2001). Sequential learning in nonhuman primates.

Trends in Cognitive Science, 5, 539–546.

Cooper, R. P., & Shallice, T. (2006). Hierarchical schemas and goals in the control of

sequential behavior. Psychological Review, 113, 887–916.

Corbetta, D., & Thelen, E (1999). Lateral biases and fluctuations in infants’ spontaneous

arm movements and reaching. Developmental Psychobiology, 34 (4), 237-255.

Cox, R. F. A., & Smitsman, A. W. (2006). Action planning in young children’s tool use.

Developmental Science, 9, 628–641.

Csibra, G. (2008). Goal attribution to inanimate agents by 6.5-month-old infants.

Cognition, 107, 705-717.

Claxton, L. J., Keen, R., & McCarty, M. E. (2003). Evidence of motor planning in infant

reaching behavior. Psychological Science, 14, 354–356.

Clearfield, M. W., & Thelen, E. (2001). Stability and flexibility in the acquisition of

skilled movement. In C. A. Nelson & M. Luciana (Eds.), Handbook of developmental cognitive

neuroscience (pp. 253–266). Cambridge, MA: MIT Press.

Cleeremans, A., & McClellend, J. L. (1991). Learning the structure of event sequences.

Journal of Experimental Psychology: General, 120, 235–253.

de Kleijn, R., Kachergis, G. & Hommel, B. (2014) Everyday robotic action: lessons from

human action control. Frontiers in Neurorobotics, 8, 13.

Dell, G. S., Burger, L. K., & Svec, W. R. (1997). Language production and serial order:

A functional analysis and a model. Psychological Review, 104, 123-147.


Diamond, A. (2002). Normal development of prefrontal cortex from birth to young

adulthood: Cognitive functions, anatomy, and biochemistry. In D. T. Stuss & R. T. Knight

(Eds.), Principles of frontal lobe function (pp. 466–503). London: Oxford University Press.

Diamond, A., Barnett, W. S., Thomas, J., & Munro, S. (2007). Preschool program

improves cognitive control. Science, 381, 1387–1388.

Dutzi, I. B., & Hommel, B. (2009). The microgenesis of action-effect binding.

Psychological Research, 73, 425-435.

Eenshuistra, R. M.,Weidema, M. A., & Hommel, B. (2004). Development of the

acquisition and control of action–effect associations. Acta Psychologica, 115, 185-209.

Elsner, B., & Hommel, B. (2001). Effect anticipation and action control. Journal of

Experimental Psychology: Human Perception and Performance, 27, 229-240.

Elsner, B., & Hommel, B. (2004). Contiguity and contingency in action-effect learning.


Elsner, B. (2007). Infants’ imitation of goal-directed actions: The role of movements and

action effects. Acta Psychologica, 124, 44–59.

Elsner, B., Hauf, P., & Aschersleben, G. (2007). Imitating step by step: a detailed

analysis of 9- to 15-month olds’ reproduction of a 3-step action sequence. Infant Behavior &

Development, 30, 325-335.

Estes, W. K. (1972). An associative basis for coding and organization in memory. In A.

W. Melton & E. Martin (Eds.), Coding processes in human memory. Washington, DC: Winston.

Fabbri-Destro, M., & Rizzolatti, G. (2008). Mirror neurons and mirror systems in

monkeys and humans. Physiology, 23, 171-9.


Falck-Ytter, T., Gredebäck, G., & von Hofsten, C. (2006). Infants predict other people's

action goals. Nature Neuroscience, 9, 878-879.

Falck-Ytter, T. (2008). Face inversion effects in autism: A combined looking time and

pupillometric study. Autism Research, 1, 297-306.

Fawcett, C., & Gredebäck, G. (2013). Infants use social context to bind actions together

into a collaborative sequence Developmental Science, 16, 841-849.

Foster, S. M., Kisley, M. A., Davis, H. P., Diede, N. T., Campbell, A. M., & Davalos, D.

B. (2013). Cognitive function predicts neural activity associated with pre-attentive temporal

processing. Neuropsychologia, 51, 211-219.

Fuster, J. (2002). Physiology of executive functions: The perception–action cycle. In:

Stuss, D. T. & Knight, R. T. (Eds.). Principles of frontal lobe function. (pp. 96–108). New York:

Oxford University Press.

Fuster, J. M. (1993). Frontal lobes. Current Opinion in Neurobiology, 3, 160–165.

Gerbing, D. W., Ahadi, S. A. & Patton, J. H. (1987).Toward a conceptualization of

impulsivity: Components across the behavioral and self-report domains. Multivariate Behavioral

Research, 22, 357–379.

Gergely, G., & Csibra. G. (2003). Teleological reasoning in infancy: The one-year-olds’

naive theory of rational action. Trends in Cognitive Science, 7(7), 287-292.

Gervain, J., Berent, I., & Werker, J.F. (2012). Binding at birth: the newborn brain detects

identity relations and sequential position in speech. Journal of Cognitive Neuroscience, 24, 564-

574.

Goren, C. C., Sarty, M., & Wu, P. Y. (1975). Visual following and pattern discrimination

of face-like stimuli by newborn infants. Pediatrics, 56, 544–549.


Gredebäck, G., Stasiewicz, D., Falck-Ytter, T., Rosander, K., & von Hofsten, C. (2009)

Action Type and Goal Type Modulate Goal-Directed Gaze Shifts in 14-Month-Old Infants.

Developmental Psychology, 4, 1190-1194.

Gredebäck, G., & Melinder, A. (2010). Infants’ understanding of everyday social

interactions: A dual process account. Cognition, 114, 197-206.

Greenwald, A. G. (1970). Sensory feedback mechanisms in performance control: With

special reference to the ideomotor mechanism. Psychological Review, 77, 73-99.

Hauf, P. (2007). Infant’s perception and production of intentional actions. In C. von

Hofsten & K. Rosander (Eds.), Progress in Brain Research: From Action to Cognition, 164,

285-301.

Hauf, P., & Aschersleben, G. (2008). Action-effect anticipation in infant action control.


Henderson, A. M. E., Wang, Y., Matz, L. E., & Woodward, A. L. (2013). Active

experience shapes 10-month-old infants’ understanding of collaborative goals. Infancy, 18, 10–

39.

Henderson, A. M. E., & Woodward, A. L. (2011). ‘Let’s work together’: what do infants

understand about collaborative goals? Cognition, 121 , 12–21.

Henry, F. M., & Rogers, D. E. (1960). Increased response latency for complicated

movements and a ‘‘memory drum’’ theory of neuromotor reaction. Research Quarterly, 31, 448–

458.

Henson, R. (1998). Short-term memory for serial order: The start-end model. Cognitive

Psychology, 36, 73-137.


Herbort, O., & Butz, M. V. (2012). Too good to be true? Ideomotor theory from a

computational perspective. Frontiers in Psychology, 3, 494.

Herwig, A., & Horstmann, G. (2011). Action-effect associations revealed by eye

movements. Psychonomic Bulletin & Review, 18, 531-537.

Hess, E. H., & Polt, J. (1960), Pupil size as related to interest value of visual stimuli,

Science, 132, 149–150.

Houghton, G., & Tipper, S. P. (1996). Inhibitory mechanisms of neural and cognitive

control: applications to selective attention and sequential action. Brain and Cognition, 30, 20-

43.

Hommel, B. (1996). The cognitive representation of action: Automatic integration of

perceived action effects. Psychological Research, 59, 176-186.

Hommel, B., Müsseler, J., Aschersleben, G. & Prinz, W. (2001). The theory of event

coding (TEC): A framework for perception and action planning. Behavioral and Brain Sciences,

24, 849 – 937.

Hommel, B., & Elsner, B. (2009). Acquisition, representation, and control of action. In E.

Morsella, J. A. Bargh, & P. M. Gollwitzer (eds.), Oxford handbook of human action (pp. 371-

398). New York: Oxford University Press.

Huestegge, L., & Kreutzfeldt, M. (2012). Action effects in saccade control. Psychonomic

Bulletin and Review, 19, 198-203.

Hull, C. L. (1931). Goal attraction and directing ideas conceived as habit phenomena.

Psychological Review, 38, 487-506.

Hupe, J. M., Lamirel, C., & Lorenceau, J. (2009). Pupil dynamics during bistable motion

perception. Journal of Vision 9, 1–19.


James, W. (1890). The principles of psychology. New York: Macmillan/Harvard

University Press.

Jackson, I., & Sirois, S. (2009). Infant cognition: Going full factorial with pupil dilation.

Developmental Science, 12, 670-679.

Johnson, M. H., Dziurawiec, S., Ellis, H., & Morton, J. (1991). Newborns’ preferential

tracking of face-like stimuli and its subsequent decline. Cognition, 40, 1–19.

Kahneman, D., & Beatty, J. (1966). Pupil diameter and load on memory. Science, 154,

1583-1585.

Kahneman, D. (1973). Attention and effort. Englewood Cliffs, New York: Prentice-Hall.

Kachergis, G., Wyatte, D., O'Reilly, R. C., De Kleijn, R., & Hommel B. (2014), A

continuous time neural model for sequential action. Philosophical Transactions of the Royal

Society of London Series B: Biological Sciences, 369, 1655.

Kayed, N. S., & Van der Meer, A. L. H. (2000). Timing strategies used in defensive

blinking to optical collisions in 5- to 7-month-old infants. Infant Behavior and Development, 23,

253–270.

Kenward, B. (2010). 10-month-olds visually anticipate an outcome contingent on their

own action. Infancy, 15, 337–361.

Kiesel, A., & Hoffmann, J. (2004). Variable action effects: Response control by context-

specific effect anticipations. Psychological Research., 68, 155-162.

Kim, R., Seitz, A. R., Feenstra, H., Shams, L. (2009). Testing assumptions of statistical

learning: Is it long-term and implicit? Neuroscience Letters, 461,145-149.

Kiraly, I., Jovanovic, B., Prinz, W., Aschersleben, G. & Gergely, G. (2003). The early

origins of goal attribution in infancy. Consciousness and Cognition, 12, 752-769.


Kirkham, N. Z., Slemmer, J. A., & Johnson, S. P. (2002). Visual statistical learning in

infancy: Evidence for a domain general learning mechanism. Cognition, 83, B35-B42.

Klein, A., Hauf, P., & Aschersleben, G. (2006). The role of action effects in 12-month

olds’ action control: A comparison of televised model and live model. Infant Behavior and

Development, 29, 535-544.

Koch, I., Keller, P. E., & Prinz, W. (2004). The ideomotor approach to action control:

Implications for skilled performance. International Journal of Sport and Exercise Psychology, 2,

362-375.

Kray, J., Eenshuistra, R., Kerstner, H., Weidema, M., & Hommel, B. (2006). Language

and action control: The acquisition of action goals in early childhood. Psychological Science, 17,

737-741.

Kunde, W. (2003). Temporal response–effect compatibility. Psychological Research, 67,

153–159.

Laeng1, B., Sirois, S., & Gredebäck, G. (2012) Pupillometry: A Window to the

Preconscious? Perspectives on Psychological Science, 7, 18-27.

Lashley, K. S. (1951). The problem of serial order in behavior. In L. A. Jeffress (Ed.),

Cerebral mechanisms in behavior. New York: Wiley.

Lewkowicz, D. J. (2004). Serial order processing in human infants and the role of

multisensory redundancy. Cognitive Processing, 5, 113–122.

Lewkowicz, D. J. (2008). Perception of dynamic & static audiovisual sequences in 3- and

4-month-old infants. Child Development, 79, 1538-1554.

Li, K. Z. H., Lindenberger, U., Rünger, D., & Frensch, P. A. (2000). The role of

inhibition in the regulation of sequential action. Psychological Science, 11, 343-347.


Lukowski, A. F., Wiebe, S. A., Haight, J. C., Waters, J. M., Nelson, C. A., & Bauer, P. J.

(2005). Forming a stable memory representation in the first year of life: Why imitation is more

than child's play. Developmental Science, 8, 279-298.

Luria, A. R. (1973). The frontal lobes and the regulation of behavior. In K. H. Pribram

(Ed.), Psychophysiology of the frontal lobes (pp. 3–26). Oxford, UK: Academic Press.

Mandler, J. M., & McDonough, L. (1995). Long-term recall of event sequences in

infancy. Journal of Experimental Child Psychology, 59, 457-474.

Mäntylä, T., Carelli, M. G., & Forman, H.(2007). Time monitoring and executive

functioning in children and adults. Journal of Experimental Child Psychology, 96, 1-19.

Marcus, G. F., Fernandes, K. J., & Johnson, S. P. (2007). Infant rule learning facilitated

by speech. Psychological Science, 18, 387-391.

Marcovitch, S., & Lewkowicz, D. J. (2009). Sequence learning in infancy: The

independent contributions of conditional probability and pair frequency information.

Developmental Science, 12, 1020-1025.

McCarty, M. E., Clifton, R. K., Ashmead, D. H., Lee, P., & Goubet, N. (2001) How

infants use vision for grasping objects. Child Development, 72, 973-987.

McCarty, M. E., Clifton, R. K., & Collard, R. R. (1999). Problem solving in infancy: The

emergence of an action plan. Developmental Psychology, 35, 1091–1101.

McCormack, T., & Atance, C. M. (2011). Planning in young children: A review and

synthesis. Developmental Review, 31, 1-31.

Michel, G. F. (1998). A lateral bias in the neuropsychological functioning of human

infants. Developmental Neuropsychology, 14, 445-469.


Montare, A. (1977). Human temporal behavior and discrimination reversal learning.

Pavlovian Integrative Physiological and Behavioral Science, 12, 232–246.

Müsseler, J., & Prinz, W. (2002) (Hrsg.). Lehrbuch Allgemeine Psychologie. Heidelberg:

Spektrum Akademischer Verlag.

Norman, D. A., & Shallice, T. (1986). Attention to action: willed and automatic control

of behaviour. In R. J. Davidson, G.E. Schwartz and D. Shapiro (Eds.) Consciousness and Self

Regulation: Advances in Research, Vol. IV pp1-18, New York: Plenum.

Olofson, E. L., & Baldwin D. (2011). Infants recognize similar goals across dissimilar

actions involving object manipulation. Cognition, 118, 258-264.

Paulus, M. (2012). Action mirroring and action understanding: An ideomotor and

attentional account. Psychological Research, 76, 760-767.

Paulus, M., Hunnius, S., & Bekkering, H. (2013). Neurocognitive mechanisms

subserving social learning in infancy: Infants’ neural processing of the effects of others’ actions.

Social Cognitive and Affective Neuroscience, 7, 774-749.

Paulus, M., Hunnius, S., van Elk, M., & Bekkering, H. (2012). How learning to shake a

rattle affects 8-month-old infants’ perception of the rattle’s sound: Electrophysiological evidence

for action-effect binding in infancy. Developmental Cognitive Neuroscience, 2, 90-96.

Paulus, M. (2014). How and why do infants imitate? An ideomotor approach to social

and imitative learning in infancy. Psychonomic Bulletin & Review, published online 28th of

February 2014.

Piaget, J. (1936, 1963). The origins of intelligence in children. New York: W.W. Norton

& Company, Inc.


Picozzi, M., de Hevia, M. D., Girelli, L., & Macchi-Cassia, V. (2010). Seven-month-old

infants detect ordinal numerical relationships within temporal sequences. Journal of

Experimental Child Psychology, 107, 359-367.

Posner, M. I., & Cohen, Y. A. (1984). Components of visual orienting. In H. Bouma &

D. G. Bouwhuis (Eds.), Attention and performance (pp. 531–554). Hillsdale, NJ: Erlbaum.

Prinz, W. (1990). A common coding approach to perception and action. In Neumann O.,

Prinz W. (Eds.), Relationships between perception and action (pp. 167-201). Berlin: Springer.

Prinz, W. (1997). Perception and action planning. European Journal of Cognitive

Psychology, 9, 129-154.

Romberg, A. R., & Saffran, J. R. (2013). Expectancy learning from probabilistic input by

Infants. Frontiers in Psychology, 3, 610.

Rosenbaum, D. A. (1987). Successive approximations to a model of human motor

programming. Psychology of Learning and Motivation, 21, 153–182.

Rosenbaum, D. A., Marchak, F., Barnes, H. J., Vaughan, J., Slotta, J., & Jorgensen, M.

(1990). Constraints for action selection: overhand versus underhand grips. In: Jeannerod, M.

(ed.), Attention and performance XIII (pp. 321–342). Hillsdale, NJ: Erlbaum.

Rumelhart, D. E., & Norman, D. (1982). Simulating a skilled typist: A study of skilled

cognitivemotor performance. Cognitive Science, 6, 1–36.

Saffran, J. R., Johnson, E. K., Aslin, R. N., & Newport, E. L. (1999). Statistical learning

of tone sequences by human infants and adults. Cognition, 70, 27–52.

Sakai, K., Kitaguchi, K., & Hikosaka, O. (2003). Chunking during human visuomotor

sequence learning. Experimental Brain Research, 152, 229–242.


Sato, A., & Itakura, S. (2013). Intersubjective action-effect binding: Eye contact

modulates acquisition of bidirectional association between our and others' actions. Cognition,

127, 383-390.

Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime user’s guide. Pittsburgh:

Psychology Software Tools Inc.

Scerif, G., Karmiloff-Smith, A., Campos, R., Elsabbagh, M., Driver, J., & Cornish, K.

(2005). To look or not to look? Typical and atypical development of oculomotor control. Journal

of Cognitive Neuroscience, 4, 591-604.

Senju, A., & Csibra, G. (2008). Gaze following in human infants depends on

communicative signals. Current Biology, 18, 668-671.

Shimpi, P., Akhtar, N., & Moore, C. (2013). Toddlers’ imitative learning in interactive

and observational contexts: The role of age and familiarity of the model. Journal of

Experimental Child Psychology, 116, 309-329.

Shin, Y. K., Proctor, R. W., & Capaldi, E .J. (2010). A review of contemporary

ideomotor theory. Psychological Bulletin, 136, 943-974.

Snodgrass, G., & Vanderwart, M. (1980). A standardized set of 260 pictures: norms for

naming agreement, familiarity, and visual complexity. Journal of Experimental Psychology:

Human Learning and Memory, 6, 174-215.

Sternberg, S., Monsell, S, Knoll, R. L., & Wright, C. E. (1978). The latency and duration

of rapid movement sequences: Comparisons of speech and typing. In G. E. Stelmach (Ed.)

Information Processing in Motor Control and Learning. New York: Academic Press. Pp. 117-

152.


Stocker, C., & Hoffmann, J. (2004) The ideomotor principle and motor sequence

acquisition: tone effects facilitate movement chunking. Psychological Research, 68, 126–137.

Suanda, S., Tompson, W., & Brannon, E. M. (2008). Changes in the ability to detect

ordinal numerical relationships between 9 and 11 months of age. Infancy, 13, 308–337.

Teinonen, T., Fellmann, V., Näätänen, R., Alku, P., & Huotilainen, M. (2009). Statistical

language learning in neonates revealed by event-related brain potentials. BMC Neuroscience, 10,

21.

Tomasello, M. (1999). The cultural origins of human cognition. Cambridge, MA:

Harvard University Press.

van den Broek, P. (1997). Discovering the cement of the universe: The development of

event comprehension from childhood to adulthood. In P. van den Broek, P. Bauer, & T. Bourg

(Eds.), Developmental spans in event comprehension and representation: Bridging fictional and

actual events (pp. 321-342). Hillsdale, NJ: Erlbaum.

Van Hof, P., Van der Kamp, J., & Savelsbergh, G. J. P. (2006). Three- to eight-month-old

infants’ catching under monocular and binocular vision. Human Movement Science, 25, 18–36.

Verschoor, S. A., Weidema, M., Biro, S., & Hommel, B. (2010). Where do action goals

come from? Evidence for spontaneous action-effect binding in infants. Frontiers in Cognition, 1,

201.

Verschoor, S. A., Eenshuistra, R. M., Kray, J., Biro, S., & Hommel, B. (2012). Explicit

learning of arbitrary and non-arbitrary action–effect relations in adults and 4-year-olds. Frontiers

in Psychology, 2, 354.

Verschoor, S. A., & Biro, S. (2012). Primacy of information about means selection over

outcome selection in goal attribution by infants. Cognitive Science, 36, 714–725.


Verschoor, S. A., Spapé, M., Biro, S., & Hommel, B. (2013). From outcome prediction to

action selection: Developmental change in the role of action-effect bindings. Developmental

Science, 16, 801-814.

Wang, Q., Bolhuis, J., Rothkopf, C. A, Kolling, T., Knopf, M., & Triesch, J. (2012).

Infants in Control: Rapid Anticipation of Action Outcomes in a Gaze-Contingent Paradigm.

PLoS ONE, 7(2).

Willatts, P. (1999). Development of means-end behavior in young infants: Pulling a cloth

to retrieve a distant object. Developmental Psychology, 35, 651–667.

Wolfe, C. D., & Bell, M. A. (2007). The integration of cognition and emotion during

infancy and early childhood: Regulatory processes associated with the development of working

memory. Brain and Cognition, 65, 3–13.

Woodward, A. L. (1998). Infants selectively encode the goal object of an actor's reach.

Cognition, 69, 1-34.

Woodward, A. L., & Sommerville, J. A. (2000). Twelve-month-old infants interpret

action in context. Psychological Science, 11, 73-76.

Woodward, A. L., Sommerville, J. A., Gerson, S., Henderson, A. M. E., & Buresh, J. S.

(2009). The emergence of intention attribution in infancy. In B. Ross (Ed.) The Psychology of

Learning and Motivation, Volume 51. Academic Press.

Wentworth, N., Haith, M. M., & Hood, R. (2002). Spatiotemporal regularity and

interevent contingencies as information for infants’ visual expectations. Infancy, 3, 303 – 321.

Zélanti, P., & Droit-Volet, S. (2011). Cognitive abilities explaining age-related changes

in time perception of short and long durations. Journal of Experimental Child Psychology, 109,

143–157.


Ziessler, M. (1994). The impact of motor responses on serial pattern learning.


Ziessler, M. (1998). Response-effect learning as a major component of implicit serial

learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 24, 962–978.

Ziessler, M., & Nattkemper, D. (2001). Learning of event sequences is based on

response–effect learning: Further evidence from a serial reaction task. Journal of Experimental

Psychology: Learning Memory and Cognition, 27, 595–613.


FIGURES

Goal

A1 A2

Goal

A1 A2

Goal

A1 A2

A B C

Figure 1. Models of sequential action. :Schematic representation of activation in A:

Chaining models of sequential action, activation cascades forward through the different

elementary actions, B: Concurrent models of sequential action, all elementary actions are

activated simultaniously where after competion through inhibition ensures the correct order of

execution, C: Integrated models of sequential action, the sequence of actions has been integrated

into a new elementary action.



T2

T3

T4

TIME

1 or 2

3 or 4

TIMET1

T2

T3

3 or 4


T2

T3

T4

TIME

1 or 2

3 or 4

TIMET1

T2

T3

3 or 4

Figure 2.

Acquisition trial.: Each trial starts with an intertrial interval of 500 ms. T1: A fixation

dot is displayed at screen center. T2: After successful fixation, faces appear at either side of the

screen where they started to pulsate. T3: Depending on the saccade target, the face at the other

side disappears and an effect sound is played for 200 ms. T4: Depending on which side was

chosen the face moves up or down whereafter a second saccade is made and a second sound

effect is played for 200 ms.

Test trial. Each trial starts with an intertrial interval of 500 ms. T1: A fixation dot is

displayed at screen center. After succesful fixation one of the previous action effects is played.

T2: The dot diasppears whereafter the same face appears on both sides. T3: The participant

freely chooses where to saccade.


Figure 3. Mean reaction times (+SE) for 9-month-olds (N=14) and 12-month-olds (N =

16) in congruent and incongruent test trials.



stimulus-locked.


response locked.


Figure 6. Vertical distance from midline toward cued direction of gaze position in pixels

for congruent and incongruent responses. Time is 0, is the moment the effect starts.


Table 1. The action sequences learned during acquisition and the congruent and

incongruent responses during test.

ACQUISITION TEST

SEQUENTIAL ACTION CONGRUENT INCONGRUENT

R1A à E1A à R2A à E2a

And

R1B à E1B à R2B à E2B

E2A à R1A

And

E2B à R1B

E2A à R1B

And

E2B à R1A

Table 2. Mean scores of acquisition phase (standard deviation in brackets).

AGE GROUP SCORES

Percentage of completed acquisition

trials

Percentage of left responses RT in ms RT

left RT

right RT up RT down


42.8 (38)

1007 (81)

999 (107)

1046 (90)

1008 (116)

1037 (83)

12-month-olds 92.0 (14)

38.5 (40)

1023 (83)

1089 (167)

982 (73)

982 (78)

1089 (164)


Table 3. Mean frequency and RT scores of test phase (standard deviation in brackets).

AGEGROUP SCORES

Percentage completed test trials

Percentage left

responses

Percentage congruent responses

ITI ms

RT ms

RT Congruent

ms

RT Incongruent

ms


43.2 (38)

47.5 (8)

1637 (323)

431 (90)

440 (104)

424 (91)

12-month- olds

96.5 (10)

29.0 (21)

49.3 (7)

1563 (393)

447 (74)

468 (83)

425 (83)

the developing cognitive substrate of sequential action ...bernhard-hommel.eu/verschoor et al. (in...

Documents